3-dimensional multivertex reconstruction from 2-dimensional tracks observations using likelihood inference
Let be vertices in the -space, each vertex producing several tracks (straight lines) emanating from it within a narrow cone with a small angle about a fixed direction (-axis). Each track is detected (by drift chambers or other detectors) by its projections on and views independently with small errors. An automated method is suggested for the reconstruction of vertices from noisy observations of the tracks projections. The procedure is based on the likelihood inference for mixtures. An illustrative...